Pricing¶
tobira is open source and free to use. For organizations that need enterprise-grade features, we offer commercial plans.
Plans¶
Available Now¶
| Community | |
|---|---|
| Price | Free |
| License | MIT |
| MTA plugins | All (rspamd, SpamAssassin, Haraka, Postfix milter) |
| Inference backends | All (FastText, BERT, ONNX, Ollama, LLM API, Two-Stage, Ensemble) |
| CLI tools | Full (init, doctor, monitor, train, evaluate, demo, distill, hub-push/pull) |
| Single-tenant deployment | |
| Docker / Kubernetes deployment | |
| Web dashboard | |
| A/B testing | |
| Active learning | |
| Knowledge distillation | |
| AI-generated text detection |
Planned¶
| Enterprise | Cloud | |
|---|---|---|
| Price | Contact us | Contact us |
| License | Commercial | Commercial |
| Includes all Community features | ||
| Multi-tenant management | Planned | Planned |
| RBAC + audit logs | Planned | Planned |
| OpenTelemetry integration | Planned | Planned |
| Grafana dashboard templates | Planned | Planned |
| SSO (SAML / OIDC) | Planned | Planned |
| Cloud-based model training | Planned | |
| Managed GPU resources | Planned | |
| Priority support + SLA | Planned | Planned |
Community (Free)¶
The open-source core includes everything you need to deploy ML-powered spam detection:
- All MTA plugins for major mail servers (rspamd, SpamAssassin, Haraka, Postfix milter)
- All inference backends with full functionality (FastText, BERT, ONNX, Ollama, LLM API, Ensemble)
- CLI-driven guided setup (
tobira init), diagnostics (tobira doctor), and monitoring - Docker Compose and Kubernetes deployment with health checks
- Web dashboard for real-time monitoring
- GDPR-aware PII anonymization for training data
- A/B testing framework for model comparison
- Active learning with uncertainty-based sampling
- Knowledge distillation (teacher → student model compression)
- AI-generated text detection
- HuggingFace Hub integration for model sharing
- Community support via GitHub Issues and Discussions
Enterprise¶
For organizations running tobira across multiple teams or domains. All features below are planned for 2026 H2 — see our Roadmap for details.
- Multi-tenant management — Isolated configurations per tenant with centralized administration
- RBAC + audit logs — Role-based access control with full audit trail
- Observability — OpenTelemetry metrics export and pre-built Grafana dashboards
- SSO integration — SAML and OIDC for enterprise identity providers
- Priority support — Dedicated support channel with SLA
Cloud¶
Hybrid cloud model — inference stays on-premise, training runs on our GPU infrastructure. All features below are planned for 2027 — see our Roadmap for details.
- Cloud-based training — Upload anonymized data, train models on managed GPUs, download results
- Managed infrastructure — No GPU procurement or maintenance required
- Includes all Enterprise features
Waitlist¶
Enterprise and Cloud plans are in the planning stage. Join the waitlist to get early access and help shape the product.
What happens next?
- Fill out a short form (email, organization, deployment scale)
- We'll reach out to understand your requirements
- Early waitlist members get priority access and input on feature prioritization
FAQ¶
Is the Community edition production-ready?¶
Yes. The Community edition includes all inference backends, MTA plugins, CLI tools, and Docker deployment — everything needed to run tobira in production for a single organization.
What's the difference between Enterprise and Cloud?¶
Enterprise is a self-hosted solution with multi-tenant management and enterprise integrations. Cloud adds managed GPU infrastructure for model training — you send anonymized data, we handle the compute.
Will pricing change?¶
Pricing details will be finalized based on early adopter feedback. Waitlist members will be notified before any changes and offered founding-member pricing.
Can I try Enterprise features before committing?¶
We plan to offer a trial period once Enterprise features are available. Join the waitlist to be notified when trials become available.